AI is moving fast. And the infrastructure behind it has to keep up.

But one thing is clear: water is at the heart of it. You need water to produce chips, to power them and to cool them. There is no AI without water.

And today, that matters more than ever. By 2030, AI could require the annual power needs of India and the annual drinking water needs of the United States.

AI is driving one of the largest infrastructure buildouts in history. More than $500 billion was invested in 2025, and that figure is expected to approach $1 trillion within the next few years. A new data centre comes online nearly every week and a new semiconductor fab every month.

Capacity alone won't be enough. Efficiency, transparency and community impact will matter just as much.

The world is right to ask tough questions about how data centres are built. Water and energy are local. So is the impact. Communities want to know how this infrastructure is being built, operated and whether it can grow without putting more pressure on local resources.

So, the answer is not to slow AI. It is to create data centres that use less, reuse more and deliver stronger performance.

The old way doesn’t work anymore. The old model wasn’t built for AI density. As AI workloads become more powerful, heat density increases and traditional cooling approaches quickly reach their limits. Yesterday’s data centres were designed around a linear model: more water in, more energy in and more strain out.

The good news is a better model already exists.

The new way is different. Less water. Less energy. Less strain on communities. We’re moving from a linear system to what nature has done for millions of years: using water in a circular way.

In fact, some next-generation data centres already use less water than a busy car wash. That is possible because technology is changing how these facilities are designed and operated.

And it all starts with liquid cooling.

When you use direct liquid cooling on chips, you get better performance and create the opportunity to use water differently. Combined with an optimised end-to-end cooling platform, it also reduces the energy needed for cooling. That’s the power of bringing CoolIT and Ecolab together.

At Ecolab, for more than 100 years we’ve helped customers grow while protecting people and natural resources for generations to come.

Today, that expertise positions us to help support AI growth across the full value chain.

With Ovivo, we added ultra-pure water to produce the most advanced chips. With CoolIT, we now bring direct liquid cooling to power the most demanding AI data centres while moving toward a near-zero water footprint. Together, we support the full AI value chain: from chips, to power, to cooling.

There is a better way to build the AI data centres of the future.

Data centres built around circular systems that use less, reuse more and deliver stronger performance.

Communities don't have to choose between growth and impact. They can benefit from AI investment while reducing the impact on people and nature. Done the right way, AI data centres don't strain communities, they strengthen them.

AI is moving fast. Now it is up to all of us to make sure the infrastructure behind it evolves just as quickly.

That’s why no one is better positioned that Ecolab—the world’s water company—to help unleash AI growth the right way.

We’re just getting started. And the best is yet to come.